PRAME Is a Golgi-Targeted Protein That Associates with the Elongin BC Complex and Is Upregulated by Interferon-Gamma and Bacterial PAMPs

Frances R. Wadelin, Joel Fulton, Hilary M. Collins, Nikolaos Tertipis, Andrew Bottley, Keith A. Spriggs, Franco H. Falcone, David M. Heery* School of Pharmacy, Centre for Biomolecular Sciences, University of Nottingham, University Park, Nottingham, United Kingdom

Abstract Preferentially expressed antigen in melanoma (PRAME) has been described as a cancer-testis antigen and is associated with leukaemias and solid tumours. Here we show that PRAME gene transcription in leukaemic lines is rapidly induced by exposure of cells to bacterial PAMPs (pathogen associated molecular patterns) in combination with type 2 interferon (IFNc). Treatment of HL60 cells with lipopolysaccharide or in combination with IFNc resulted in a rapid and transient induction of PRAME transcription, and increased association of PRAME transcripts with polysomes. Moreover, treatment with PAMPs/IFNc also modulated the subcellular localisation of PRAME proteins in HL60 and U937 cells, resulting in targeting of cytoplasmic PRAME to the Golgi. Affinity purification studies revealed that PRAME associates with Elongin B and Elongin C, components of E3 ubiquitin complexes. This occurs via direct interaction of PRAME with Elongin C, and PRAME colocalises with Elongins in the Golgi after PAMP/IFNc treatment. PRAME was also found to co-immunoprecipitate core histones, consistent with its partial localisation to the nucleus, and was found to bind directly to histone H3 in vitro. Thus, PRAME is upregulated by signalling pathways that are activated in response to infection/inflammation, and its product may have dual functions as a histone-binding protein, and in directing ubiquitylation of target proteins for processing in the Golgi.

Citation: Wadelin FR, Fulton J, Collins HM, Tertipis N, Bottley A, et al. (2013) PRAME Is a Golgi-Targeted Protein That Associates with the Elongin BC Complex and Is Upregulated by Interferon-Gamma and Bacterial PAMPs. PLoS ONE 8(2): e58052. doi:10.1371/journal.pone.0058052 Editor: William R. Abrams, New York University, United States of America Received January 5, 2012; Accepted January 30, 2013; Published February 27, 2013 Copyright: ß 2013 Wadelin et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was funded by Leukaemia Lymphoma Research via a clinical training fellowship to FRW (07064). JF was funded by the Royal Pharmaceutical Society of Great Britain via an Academic Excellence Award to DMH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]

Introduction pathogenic [16], although until recently there has been little insight into its functions in mammalian cells. Nuclear PRAME/MAPE/OIP4 is an atypical cancer-testis antigen localised PRAME has been implicated in transcriptional repres- whose expression is associated with leukaemias and a large sion via association with retinoic acid receptor complexes [17]. proportion of solid tumours (reviewed in [1]). Unlike other More recently, while this study was in progress, PRAME was cancer-testis antigens whose expression is restricted to testis, the shown to be a chromatin-associated protein enriched at nuclear PRAME gene shows low level expression in other normal tissues factor Y (NFY) target genes, in association with Elongin and including endometrium, ovary and placenta [2]. The restricted Cullin-2 proteins [18]. However, a large proportion of endog- expression pattern of PRAME in normal tissues and its enous PRAME protein is observed in the cytoplasmic compart- overexpression in tumours renders it a useful marker of minimal ment in different cell lines [1,19]. Given its reported interaction residual disease after chemotherapy, and an attractive target for with bacterial outer membrane proteins [16] and the rapid immunotherapy, particularly in acute myeloid leukaemia (AML) evolution of the PRAME multigene family in humans [15], and chronic myeloid leukaemia (CML) [2,3,4,5,6,7,8,9,10,11,12]. similar to the Nacht, LRR, PYD domain (NALP) gene family In PRAME-negative leukaemias the gene can be induced by [20], we hypothesised that PRAME might be regulated by demethylating agents [2,13,14], but little else is known regarding signalling pathways activated in proinflammatory responses. We pathways that regulate PRAME expression. therefore set out to investigate whether PRAME expression is PRAME is a member of a rapidly evolving multigene family, modulated by signalling molecules such as IFNc or microbial and encodes a leucine-rich repeat (LRR) protein sharing PAMPs. We also endeavoured to isolate PRAME-interacting structural similarity with Toll-like receptors [1,15]. The extensive proteins, to provide further insight into its cellular functions. duplication rate of PRAME and other cancer testis antigen genes is suggestive of roles in chemosensing, reproduction or immunity [15]. PRAME was originally identified in a yeast two-hybrid screen for proteins that bind outer membrane proteins of

PLOS ONE | www.plosone.org 1 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Results the cytoplasm of these cells within 24 hrs of transfection (Fig. 2A). Staining with a specific antibody recognising PRAME confirmed Upregulation of PRAME expression by IFNc and bacterial the identity of the recombinant PRAME-EGFP and PRAME- PAMPs FLAG constructs, in contrast to neighbouring non-transfected cells PRAME protein contains a series of leucine-rich repeats similar which showed negligible endogenous levels of PRAME (Fig. 2A). to those found in the LRR and is predicted to be In contrast to PRAME-EGFP, EGFP alone showed approximately structurally similar to human Toll-like receptors (TLRs) [1,15]. equal distribution in the nucleus and cytoplasm whereas co- Unlike the membrane-associated TLRs, PRAME is an intracel- transfected PRAME-FLAG was predominantly cytoplasmic lular protein found in both the nuclear and cytoplasmic (Fig. 2A, lower panels). These observations are consistent with compartments [1]. TLRs function in innate immunity as sensors suggestions that PRAME is likely to have both nuclear and of microbial PAMPs or other ligands, and their expression is cytoplasmic functions, although the distribution between these regulated by these molecules and also by IFNc, a proinflammatory compartments may vary in different cell types and conditions. cytokine produced in response to infection. We therefore assessed To determine whether exposure to bacterial PAMPs/IFNc had whether PRAME expression in leukaemic cells such as HL60 might any effect on the subcellular localisation of PRAME, we be modulated by exposure to lipopolysaccharide (LPS) and IFNc, performed immunofluorescence staining of cells exposed to these either as single inducing agents or in combination. HL60 agents, or controls. A longer treatment time (4 hours) was used to leukaemic cells have low levels of PRAME transcripts [1], and allow for expression of PRAME proteins in the cells. As shown in are known to express TLR2 and TLR4 [21]. Treatment of HL60 Fig. 2B, treatment of HL60 cells with LPS/IFNc (middle panels) cells with IFNc or LPS alone did not significantly alter the or PGN/IFNc (lower panels) for 4 hours resulted in localisation of expression of PRAME (as determined by RT-qPCR – data not PRAME to large cytoplasmic foci. A similar effect was observed in shown). However, as shown in Fig. 1A, a strong increase in the leukaemic cell line U937 in response to LPS/IFNc treatment PRAME transcript levels was observed within 1 hour of combined (Fig. 2C). Quantitation of PRAME-containing foci in LPS/IFNc- treatment with LPS and IFNc. This increase was transient as treated and or control cells revealed a major change in the PRAME levels were lower at 4 hours post-treatment, and restored subcellular distribution of PRAME in response to LPS/IFNc to background levels within 24 hours, suggesting that turnover of treatment (Fig. 2E&F). PRAME mRNA may be relatively rapid. The induction of PRAME The polarised distribution of the PRAME-containing foci, by LPS/IFNc was inhibited by pre-treating cells with actinomycin suggested that these structures may represent the trans Golgi D for one hour prior to addition of LPS/IFNc (Fig. 1B), indicating network, and staining of HL60 cells with the 58K Golgi marker that the increase in PRAME levels is due to de novo transcription confirmed that these structures appear to be linked to the Golgi rather than altered stability of the transcripts. network and show considerable colocalisation with PRAME To examine whether LPS/IFNc results in increased translation (Fig. 2D). We noted that the nuclei of LPS/IFNc treated HL60 of PRAME transcripts, polysome gradients were prepared from and U937 cells appeared to undergo a morphological change extracts of HL60 cells treated or untreated with LPS and IFNc.As resulting in ‘kidney’ shaped nuclei being visible after 4 hours shown in Fig. 1C, within 4 hours post-treatment with LPS/IFNc, treatment (Fig. 2B&C), and that this phenotype was prevalent in PRAME transcripts were substantially enriched in polysome many of the treated cells, but less obvious in controls (Fig. 2E). fractions, as confirmed by densitometry (lower panels). In contrast, a control transcript (b-actin) showed little change. These results PRAME and the 58K Golgi marker accumulated at the apex of indicate that upregulation of PRAME gene expression by LPS/ the invaginated region (Fig. 2D, lower panels). Thus, we conclude IFNc in HL60 cells may be achieved not only through increased that in HL60 and U937 cells, cytoplasmic PRAME proteins are transcription, but also effects on translation of PRAME transcripts. targeted to the Golgi network following exposure to LPS/IFNc. We also assessed whether PRAME expression could be induced Consistent with this, subcellular localisation prediction algorithms by other microbial PAMPs in addition to LPS. As shown in (PSORT.org) and other sequence analysis resources predict that Fig. 1A, treatment of HL60 cells with PAMPs such as muramyl PRAME is likely to be a Golgi-targeted protein [22]. Staining of dipeptide (MDP), zymosan A (ZYM) or mannan (MAN) for endogenous PRAME in MCF-7 breast cancer cells also revealed different times, either alone or in combination with IFNc, had its association with Golgi-like structures (Fig. 2G). relatively little effect on PRAME expression. In contrast, peptido- glycan (PGN) in combination with IFNc substantially increased Functional association of PRAME with Elongin BC PRAME expression, similar to the LPS/IFNc response. Induction complexes of PRAME by PGN/IFNc was also transient, with peak levels To further investigate the functions of PRAME, we performed observed within 1 hour post-treatment (Fig. 1A). Taken together, affinity purification experiments to identify PRAME-associated these results indicate that, similar to genes encoding other LRR proteins. Cleared whole cell lysates of HL60 cells were applied to proteins such as NALPs and TLRs, PRAME expression in HL60 GST-PRAME or control GST beads. After extensive washing, cells may be rapidly induced by signalling pathways activated in affinity purified proteins were visualised by silver staining, response to bacterial infection/inflammation. revealing a number of PRAME-specific bands (Fig. 3A). We selected two bands in the molecular range 10–20 kDa, which were Subcellular localisation of PRAME is altered by bacterial among the most abundant and which were clearly absent in the PAMPs/IFNc controls (Fig. 3A). Following gel excision and extraction, these Endogenous or recombinant PRAME proteins have been samples were subjected to reverse phase liquid-chromatography reported to localise to both nuclear and cytoplasmic compartments followed by high mass accuracy mass spectrometry. Polypeptides in a variety of cell types, including adherent cells [19] and were identified using the MASCOT search engine [23]. As shown leukaemic cell lines [1]. Transient transfection of U2OS or in Table 1, the top hits were the human ELB (elongin B) and ELC HEK293 cells with recombinant PRAME-EGFP or PRAME- (elongin C) proteins. These are components of the E3 ubiquitin FLAG revealed both nuclear and cytoplasmic staining, with the ligase complexes and are implicated in both protein ubiquitylation majority of signal (70–80% of total fluorescence) being observed in and transcription elongation.

PLOS ONE | www.plosone.org 2 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Figure 1. Transcriptional and translational regulation of PRAME by PAMPs/IFNc. (A) RT-qPCR measurements of PRAME gene expression relative to control GAPDH in HL60 cells in response to treatment with different PAMPs including lipopolysaccharide (LPS), peptidoglycan (PGN), muramyl dipeptide (MDP), zymosan (ZYM) and mannan (MAN) either alone (PBS) or in combination with IFNc. Numbers on the x-axis indicate the time in hrs post-treatment. qPCR quantifications were performed in triplicate and the data shown represents the mean of two independent experiments, with error bars indicating standard deviations. The data is presented as fold induction relative to levels obtained at 0 hr (baseline). (B) RT-qPCR experiment performed as in (A) showing effect of pre-treatment with actinomycin D (10 mg/ml) or PBS on induction of PRAME transcript in

PLOS ONE | www.plosone.org 3 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

HL60 cells by LPS/IFNc (1 hour). (C) Association of PRAME transcripts with polysomes in HL60 cells following treatment with LPS/IFNc for 0 and 4 hrs. Following cycloheximide treatment and sucrose density centrifugation of HL60 cell lysates, gradients were fractionated with continuous monitoring at 254 nm, to generate polysome profiles (top panel). RNA extracted from polysome fractions was analysed by Northern blotting and the PRAME transcripts visualised by phosphoimager (middle panels) and quantified by densitometry (bottom panels). b-actin was used as a control probe. doi:10.1371/journal.pone.0058052.g001

To validate these potential interactions, co-immunoprecipita- Cullin 2 (CUL2) in co-immunoprecipitation experiments using tion experiments were performed. As shown in Fig. 3B, endog- whole cell extracts from PRAME transfected HEK293 cells enous ELB and ELC proteins were successfully co-purified from (Fig. 3B). Taken together, these results provide strong evidence for HEK293 cell extracts in complex with transiently expressed functional interaction between PRAME and components of the recombinant PRAME-FLAG-His. Similarly, GST-PRAME, but Elongin/Cullin E3 complexes, consistent with the not GST alone, was able to purify both ELB and ELC from HL60 recent report from Costessi et al [18]. Our data further suggest that whole cell extracts (Fig. 3C). However, incubation of GST- this is due to direct interaction of PRAME and ELC. Immuno- PRAME with 35 [S]-labelled in vitro translated proteins revealed fluorescence staining of endogenous ELB and ELC proteins in that direct interaction occurs between PRAME and ELC, but not HL60 cells revealed their localisation to cytoplasmic structures ELB (Fig. 3D). This was confirmed using yeast two hybrid studies, resembling Golgi (Fig. 3F). Unlike PRAME, this association of which detected a strong interaction between LexA-PRAME and ELB and ELC with Golgi was not dependent on treatment with GAL4 activation domain-fused human or C. elegans ELC proteins, PAMP/IFNc (Fig. 3F). Some nuclear staining of ELB and ELC (which share 74% sequence identity), but not ELB (Fig. 3E). was also detected, consistent with proposed roles of these proteins Moreover, this interaction with ELC was disrupted by mutations in transcriptional elongation [24]. Co-staining of LPS/IFNc (L47D-L49D-Y88D-Y91D) in a conserved region required for treated HL60 cells revealed extensive overlap of ELC and binding of NLP48 [24] (Fig. 3E). No interaction was detected PRAME in Golgi-like structures, supporting the conclusion that using ELB, although western blots on yeast cell-free extracts these proteins colocalise to the Golgi (Fig. 3G). In contrast, CUL2 confirmed the expression of both ELB and ELC constructs was found to be almost exclusively nuclear under similar (Fig. 3E). We also detected the Elongin BC-associated protein conditions (Fig. 3F), which may be consistent with the identifica-

Figure 2. PRAME localises to the Golgi network following LPS/IFNc treatment. (A) HEK293 cells (upper panels) were transiently transfected with PRAME-EGFP (green) and stained with a-PRAME antibody (red) to confirm the identity of the overexpressed EGFP fusion protein. U2OS cells (lower panels) were cotranfected with GFP (green) and PRAME-FLAG (red). Merged images indicate the extent of coincidence of the EGFP and a- PRAME signals, and nuclear DNA is indicated (blue). The right hand panels are western blots showing detection of GFP or PRAME-EGFP proteins in whole cell extracts of transfected U2OS cells. (B) Immunostaining of endogenous PRAME in HL60 cells using a-PRAME antibody following treatment with PBS, LPS/IFNc or PGN/IFNc for 4 hrs. (C) Immunostaining of endogenous PRAME in U937 cells with a-PRAME following treatment with LPS/IFNc for 0, 1 and 4 hrs. (D) HL60 cells treated with LPS/IFNc for 4 hrs and immunostained with a-Golgi 58K (green) and a-PRAME (red). Merged images show the extent of colocalisation of both proteins. For immunofluorescence (A–D), nuclear DNA was stained using Hoechst 33258 and images were captured using a LSM510 confocal laser scanning microscope. (E) Immunostaining of endogenous PRAME in HL60 cells using a-PRAME antibody following treatment with PBS or LPS/IFNc for 4 hrs. (F) Quantification (n = 60) of the percentage of cells in (E) containing PRAME cytoplasmic foci in treated cells or controls. (G) Immunostaining of endogenous PRAME in MCF-7 cells using a-PRAME antibody. doi:10.1371/journal.pone.0058052.g002

PLOS ONE | www.plosone.org 4 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Figure 3. PRAME associates with the Elongin BC complex. (A) SDS-PAGE and silver staining showing affinity capture of proteins from HL60 whole cell extracts (WCE) by immobilised GST or GST-PRAME proteins. GST and GST-PRAME proteins are indicated. Putative PRAME-specific bands are indicated and bands of approximately 12 kDa and 17 kDa were excised for mass spectrometry analysis. (B) Co-immunoprecipitation of PRAME with Elongin complex components. Whole cell extracts of HEK293 cells transfected with PRAME-FLAG-6xHis (or empty vector control) applied to anti-FLAG sepharose beads as described in Materials and Methods. After extensive washing, co-purified PRAME and E3 ubiquitin ligase complex components were detected by western blotting using specific antibodies as indicated. (C) GST-pulldown experiment showing binding of ELB and ELC proteins in HL60 whole cell extracts to GST or GST-PRAME proteins. The top panel is a Coomassie-stained gel showing the input whole cell extract, and the purified GST and GST-PRAME proteins. The lower panels are western blots revealing PRAME, ELB and ELC proteins bound to GST proteins. (D) GST- pulldown experiments revealing interactions of 35 [S]-labelled in vitro translated human ELC (hELC), C.elegans ELC (wELC), C.elegans ELC (L47D-L49D- Y88D-Y91D) (wELC mutant) and C.elegans ELB proteins with GST or GST-PRAME. (E) Yeast two hybrid assays of LexA-PRAME interactions with GAL4 AD-fused human ELC (hELC) or C.elegans proteins (wELB, wELC, wELC mutant). Western blots of the HA-tagged elongin fusion proteins are also shown. Reporter activity is expressed as b-galactosidase activity normalised to amount of protein in the extracts. (F) Immunofluorescence staining showing subcellular localisation of endogenous ELC, ELB and CUL2 proteins in HL60 cells. (G) Immunofluorescence staining showing colocalisation of endogenous ELC and PRAME proteins in HL60 cells following treatment with LPS/IFNc for 4 hours. doi:10.1371/journal.pone.0058052.g003 tion of a CUL2/Elongin BC/PRAME complex in nuclear extracts Interaction of PRAME with histones [18]. Other ubiquitin ligase complexes have been implicated in In addition to Elongins, the mass spectrometry analysis also trafficking to the Golgi and ER, and it remains possible that detected histones as putative PRAME binding proteins (Table 1). cytoplasmic PRAME can associate with other such as the Nuclear localised PRAME was recently reported to act as a Golgi-targeted CUL3 or CUL7 proteins [25,26]. coregulator of gene expression, being recruited to a subset of NFY-

PLOS ONE | www.plosone.org 5 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Table 1. Mass Spectrometry Identification of PRAME Binding Proteins.

Band Proteins identified Number of peptides Mowse score

1 ELB 6 359 Histone H2B 2 73 2 ELC 6 397 Histone H4 2 108

GST-PRAME binding proteins were affinity purified as described in Materials & Methods. Bands of approximately 17 kDa and 12 kDa (Fig. 3A) were excised for mass spectrometry analysis. MS/MS fragmentation data were used to search the human NCBI database using the MASCOT search engine. Probability-based Mowse scores are shown. doi:10.1371/journal.pone.0058052.t001 regulated promoters [18]. While PRAME was demonstrated to associate with chromatin, it was reported not to interact with DNA directly, thus the mechanism by which it associates with chromatin was not established [18]. Immunoprecipitation of epitope-tagged PRAME from HEK293 whole cell extracts confirmed that all four core histones can co-immunoprecipitate with PRAME (Fig. 4A). In addition, GST-pulldown experiments using HL60 whole cell extracts (Fig. 4B) or core histone preparations (Fig. 4C) also detected interactions of PRAME with histone H3. Whole cell extracts may contain histones as nucleosomes, which could account for the co-precipitation of all four core histones from cell extracts. However core histone preparations are generally DNA- free. Therefore, the strong detection of histone H3 in core histone binding experiments (Fig. 4C) suggests that direct interaction of PRAME with H3 may enable PRAME to associate with chromatin.

Discussion Although PRAME is an important biomarker in certain haematological malignancies and solid tumours, the regulation of this gene is poorly understood. Expression of PRAME in malignant cells is known to be subject to epigenetic regulation, as we and others have shown that DNA demethylating agents can induce PRAME in tumours and leukemic cell lines [2,13,14]. However, until now, little else was known regarding the regulation of PRAME expression. In this study we provide the first evidence that PRAME may be regulated by signalling pathways involved in innate immune responses. Innate immune pathways coordinate appropriate responses to infection or immunological challenge, and also play a prominent Figure 4. Binding of PRAME to histone H3. (A) Co-immunopre- role in the regulation of haematopoiesis [27,28]. Aberrant activity of cipitation of endogenous histones with PRAME-FLAG-6xHis isolated such pathways is associated with myelodysplastic syndromes and from extracts of transfected HEK293 cells. Input lanes (left panels) show haematological malignancies [28]. For example, chronic inflam- the presence of endogenous or FLAG-tagged proteins in extracts from mation is involved in the origin and progression of certain B-cell cells transfected with PRAME-FLAG or empty vector. After IP with anti- FLAG antibody, immunoblots were performed with the antibodies lymphomas and can be mediated through TLRs and other indicated. (B) GST-pulldown experiment showing association of components of the innate immune system, that are activated histones with GST or GST-PRAME proteins. Whole cell extracts of through recognition of microbial PAMPs [29]. Similar pathways are HL60 cells were incubated with immobilised GST or GST-PRAME likely to be involved in progression of myeloid leukaemias. For proteins, and bound proteins separated by SDS-PAGE. Immunoblots example, TLR2 and TLR4 which mediate responses to LPS and were performed with specific antibodies to detect association of histones H2A, H2B, H3 and H4 with GST proteins. (C) Direct association PGN are both activated and upregulated following exposure of of histones with GST or GST-PRAME proteins. Core histone preparations HL60 cells to these agents [21]. Similarly, PGN-mediated activation were incubated with GST beads. Following extensive washing, bound of TLR2 and other pathways has been shown to enhance expression histones were separated by SDS-PAGE and revealed by western blotting of chemoattractant genes in human monocytic leukaemia cells [30]. using specific antibodies as indicated. TLR signalling converges with cytokine signalling and other doi:10.1371/journal.pone.0058052.g004 pathways to amplify or moderate immune responses. For example, the proinflammatory cytokine IFNc is produced by macrophages

PLOS ONE | www.plosone.org 6 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME following intracellular invasion and is known to prime macrophages Due to poor performance of available PRAME antibodies in to respond to infection [27]. This combination of IFNc with TLR detecting endogenous proteins by western blotting, we were activation can ensure rapid induction of subsets of proinflammatory unable to reliably detect an increase in intracellular levels of genes. Thus, given the predicted structural similarity of PRAME to soluble PRAME in HL60 cells following IFN/LPS induction. It is the extracellular LRR domains of TLR2 and TLR4 [1,15], and its possible that the detection of endogenous PRAME in cell-free reported ability to bind bacterial OPA proteins [16] we hypothesised extracts might also be hampered by its retention in the insoluble that PRAME might have some role as an intracellular sensor of fraction of cell lysates, due to increased localisation to Golgi/ER invading pathogens. Such an inducible expression profile might macromolecular structures. account for its generally low or undetectable levels in normal tissues. Using an affinity purification approach to explore the functions These observations prompted us to investigate whether PRAME of PRAME through its molecular partners, we identified an expression might be upregulated in response to microbial PAMPs interaction of PRAME with the Elongin/Cullin E3 ubiquitin ligase and/or IFNc. complex as confirmed in a range of binding assays and co- Solid tumours can also induce proinflammatory pathways, purification experiments (Fig. 3A–G). Co-immunoprecipitation producing a microenvironment that favours tumour growth and experiments detected ELB, ELC and CUL2 proteins in complex proliferation. In addition, tumour-associated leucocytes may with PRAME in HL60 cells. This association of PRAME with the produce cytokines that stimulate tumour growth and angiogenesis Elongin/Cullin complex is consistent with a recent report from [31]. This can occur through activation of TLRs, or through IFNc Costessi et al. [18], in which a similar complex was identified in production resulting in activation of downstream effectors such as K562 cells, which contain high levels of PRAME protein. nuclear factor kappa B (NFkB), interferon regulatory transcription Interestingly, these data are also consistent with a previous study factors (IRFs) and signal transducers and activators of transcription which noted an interaction between PRAME and the E2 ubiquitin (STATs) [31]. These transcription factors are implicated not only UBE2F, in a yeast two-hybrid screen [35]. In our hands, in the regulation of proinflammatory genes, but also in cancer- direct interactions appear to occur between PRAME and ELC, related inflammation and in aberrant gene regulation in tumours based on yeast two-hybrid and GST-pulldown assays (Fig. 3D&E), [31]. The mechanism by which PAMP/IFNc treatment induces which is consistent with observations for other complexes involving PRAME transcription remains to be characterised, but is likely to LRR and BC box proteins. Thus, as observed for its counterparts involve one or more of these effectors. Interestingly, the proximal in similar complexes, the function of PRAME may be to target promoter of the PRAME gene contains a number of potential substrates for ubiquitylation by the Elongin/Cullin complex, binding sites for NFkB, IRFs and STATs. Further studies will be although the identity of these potential substrates remains required to characterise the cis and trans acting factors that regulate unknown. PRAME expression. Elongins have both nuclear and cytoplasmic functions, having In addition to the observed effects on de novo transcription roles in protein ubiquitylation and gene transcription [24]. (Fig. 1A&B), treatment of HL60 cells with PAMPs/IFNc also Interestingly, staining of HL60 cells for endogenous ELC revealed appeared to increase recruitment of PRAME transcripts into it to be mainly cytoplasmic and at least partly associated with the polysomes (Fig. 1C). This suggests that PRAME may also be Golgi network (Fig. 3F&G), thus indicating that a major site of subject to translational regulation, perhaps as a consequence of the interaction of PRAME and Elongins is in the cytoplasmic extended 59UTR sequences present in PRAME transcripts [1]. It compartment. CUL2 on the other hand appeared to be almost remains to be seen whether such insights into the regulation of this exclusively nuclear (Fig. 3F). We did not observe any major change tumour biomarker will enable the manipulation of its expression in in the subcellular localisation of ELB, ELC or Cul2 following PRAME-negative tumours. As PRAME is undergoing evaluation in IFN/LPS treatments (data not shown), and antibody conflicts clinical trials as an anti-cancer vaccine target in non-small cell lung precluded co-staining of ELC with the Golgi marker. However, cancer (NCT01159964), its re-activation in tumours might the localisation of ELB and ELC to polarised cytoplasmic foci is facilitate their targeting by immunotherapeutics. consistent with Golgi/ER association. It is not yet known whether Endogenous and overexpressed PRAME proteins were ob- PRAME can associate with other Cullin complexes, such as Golgi- served in both the nucleus and cytoplasmic compartments in resident CUL3 or CUL7 proteins [25,26]. different cancer cell lines (Fig. 2A–D). In addition to the observed Our affinity purification experiment also identified a novel upregulation of PRAME expression, treatment of HL60 or U937 interaction of PRAME with core histones (Table 1). Direct leukemic cells with PAMPs/IFNc resulted in the accumulation of interaction appears to be mediated chiefly through histone H3, cytoplasmic PRAME in Golgi-like structures (Fig. 2B&C). This which shows strongest association with PRAME (Fig. 4C). This was supported by the observed colocalisation of PRAME and the result is consistent with the recent report that nuclear PRAME can 58K Golgi marker (Fig. 2D). A significant proportion of treated be recruited to chromatin and functions in gene regulation, HL60 cells were found to begin to undergo lobulation of the although it lacks DNA binding activity [18]. In addition, PRAME nucleus (Fig. 2E). Similar effects on nuclear shape have been has previously been reported to co-immunoprecipitate with the reported for a HL60 cell line undergoing retinoic acid induced histone methyltransferase EZH2 [17], which promotes trimethyla- granulopoiesis, with the Golgi juxtapositioned between nuclear tion of H3K27. Given its interaction with Elongin BC complexes, lobes [32]. Interestingly, nuclear lobulation in macrophages it is therefore conceivable that PRAME might play some role in enables them to infiltrate other tissues in response to chemoat- histone ubiquitylation, although there is currently no evidence to tractants. Consistent with reports that PRAME is expressed in support this. Several different nuclear E3 ubiquitin ligase breast tumours, [33,34] PRAME was readily detected in MCF7 complexes are known to catalyse the addition of ubiquitin to core cells, by immunofluorescence staining (Fig. 2G), where it also histones [36]. This includes the CUL2 complex which promotes showed a strong association with Golgi-like structures. The monoubiquitylation of H2B at residue K120, through the functional consequences of the association of PRAME with the interaction of the histone with the BC-box protein BAF250b Golgi remains to be determined, such as whether PRAME plays [37]. Future studies will therefore need to focus on assessing role in nuclear lobulation in macrophages or other monocytic whether depletion or overexpression of PRAME can influence cells. ubiquitylation of histones or other target proteins.

PLOS ONE | www.plosone.org 7 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

In conclusion, our study provides novel insights into the amyl dipeptide (MDP, Sigma-Aldrich), 100 mg/ml zymosan A functions and regulation of PRAME, as summarised in Figure 5. (ZYM, Sigma-Aldrich) or 10 mg/ml mannan (MAN, Sigma- We hypothesise that, like TLRs, PRAME may be upregulated in Aldrich). Actinomycin D (10 mg/ml, Sigma-Aldrich) was applied response to encounters with microbial pathogens, and may be to cells 1 hr prior to other treatments. involved in targeting intracellular PAMPs to the Golgi for U2OS and HEK293 cells were seeded onto coverslips 16 hrs ubiquitylation and processing. The activation of similar cytokine prior to transient transfection using CalPhosTM mammalian signalling pathways in cancer cells might therefore account for the transfection kit (Clontech). expression of PRAME in tumours, and its low levels in normal tissues. Finally, direct interactions of PRAME with histone H3 in Expression plasmids addition to elongins may be important in its functions in gene PRAME-EGFP and PRAME-FLAG expression vectors [19] transcription. These insights will aid future efforts to establish were gifts from P.Gailly. Yeast two hybrid pGADT7 vectors whether PRAME expression in tumours has roles in the initiation expressing C.elegans CUL2, ELB, ELC, ELC mutant (L47D/ or progression of haematological malignancies, and/or in immune L49D/Y88D/Y91D) and human ELC [38] proteins containing responses to infection. haemagluttinin (HA) epitope-tags were gifts from K.Yamanaka. A PCR fragment containing PRAME coding sequence was generated Materials and Methods from pCMV-Sport6-PRAME (Geneservice) using the following primers: 59-atggaacgaaggcgtttg-39 and 59-ctagttaggcatgaaacagggg- Cell lines, reagents and transient transfections 39. The fragment was subcloned into SmaI digested pBSKSII, to HL60, U937, U2OS, MCF-7 and HEK293 cell lines were generate pBSKSII-PRAME. pBSKSII-PRAME was digested with purchased from the European Collection of Cell Cultures BamHI and XhoI and the insert fragment was subcloned into (ECACC). HL60 and U937 leukaemic cell lines were maintained pGEX4T1, to generate GST-PRAME. This fragment was also in RPMI-1640. HEK293, U2OS and MCF-7 cells were main- subcloned into pBTM116mod and pASV3mod [35] to generate tained in Dulbecco’s modified Eagle medium. Culture medium LexA-PRAME and VP16-PRAME, respectively. PRAME-FLAG- was supplemented with 10% foetal bovine serum, 2 mM His was generated by PCR from PRAME-FLAG [19] using glutamine, 100units/ml penicillin and 100 mg/ml streptomycin primers: 59-aaaaggatccgccgccatggaacgaaggcgttt g-39 and 59- and cells were maintained at 37 C in a 5% CO incubator. Cell u 2 aaaactcgagctaatggtgatggtgatgatgcttgtcatcgtcatccttgtagta-39. The cultures were split 24 hrs prior to treatment with 100U/ml insert was subcloned into the BamHI and XhoI sites in interferon-gamma (IFNc, AbD Serotec), 50 ng/ml Salmonella pcDNA3.1(+). minnesota lipopolysaccharide (LPS, Sigma-Aldrich), 5 mg/ml Bacil- lus subtilis peptidoglycan (PGN, Sigma-Aldrich), 10 mg/ml mur-

Figure 5. Potential Functions and Regulation of PRAME. Activation of TLRs and other signalling receptors by PAMPs and cytokines associated with infections or tumours (1) results in both transcriptional (2) and translational (3) upregulation of PRAME. Cytoplasmic PRAME can associate with Elongin proteins located in the Golgi/ER network (4). Association of PRAME with intracellular PAMPs or other molecules (5) may facilitate their targeting to the Golgi for modification, secretion or destruction by Elongin/Cullin Ubiquitin (6). Interactions of PRAME proteins with Elongins, Cul2 or Histone H3 in the nucleus (7) may be involved in the regulation of gene expression. doi:10.1371/journal.pone.0058052.g005

PLOS ONE | www.plosone.org 8 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Polysome analysis and Northern blotting Cytospin 4 cytocentrifuge. Representative images of IFNc/PAMP- After cycloheximide treatment of HL60 cells and preparation of treated cells and PBS-treated controls immunostained with a- cell lysates [39], sucrose density gradients were used to separate PRAME antibody were quantified by manual scoring of the ribosomes into polysomal and subpolysomal complexes. Gradients percentage of cells (n = 60) showing large cytoplasmic PRAME- were fractionated with continuous monitoring at 254 nm, and containing foci. RNA was isolated from each fraction as described previously [40]. Northern analysis of RNA isolated from sucrose density gradients Protein affinity purification, co-immunoprecipitation and was performed as described previously [41], using 32P-labelled GST-pulldown assays probes for full-length PRAME and actin. Activity was visualised Preparation of whole cell extracts of mammalian cells was using a Storm825 phosphorimager (GE Healthcare) and Im- performed by using freeze thaw cycles in cell lysis buffer with ageQuantTL software (GE Healthcare). protease inhibitors as described previously [39]. GST-PRAME and empty pGEX-4T1 (or GST) vectors were transformed into Reverse transcription PCR and qPCR E.coli Rosetta cells and selected on LB plates containing ampicillin. RNA was extracted using EZNATM total RNA kit I (Omega Single colonies were cultured to exponential phase in LB Bio-Tek). RNA was reverse transcribed using oligo(dT)12-18 and ampicillin broth at 37uC, and GST protein expression was Superscript III (Invitrogen) according to the manufacturer’s induced by addition of, 0.5 mM isopropyl b-D-1-thiogalactopyr- instructions. RTqPCR was performed using Brilliant II SYBR anoside (Merck) for 16 hrs at 20uC. Bacterial pellets were green qPCR mix (Stratagene) and primers: PRAME: 59- resuspended in NTN (20 mM Tris pH8, 100 mM NaCl, 0.5% tgctgatgaagggacaac at-39 and 59-cag cacttgaagtttccacct-39; NP40, 1 mM DTT 1X Complete EDTA-free protease inhibitor GAPDH: 59-aggtgaaggtcggagtcaac-39 and 59-gatga- cocktail (Roche 11836170001), sonicated and then centrifuged at caagcttcccgttct-39. PCR was conducted using Mx3005P real-time 15000rpm for 40 mins at 4uC. The supernatant was added to PCR system (Stratagene) with the following conditions: hot start prewashed glutathione sepharose beads (GE Healthcare) that had activation at 95uC for 10 mins followed by 40 cycles of 95uC for been blocked with 0.5% milk and incubated for 16 hrs at 4uC with 30 secs, 59uC for 1 min. PCR reactions were performed in rotation. The beads were washed extensively in NTN buffer. triplicate for each sample. The threshold cycle number of PRAME Affinity capture was performed using 1 mg whole cell extract or was normalised to that of GAPDH. Fold increase was calculated 15 mg purified core histones (Sigma H9250). GST, GST-PRAME using the comparative (DDCt) method, as previously described and other proteins bound to beads were resolved by SDS-PAGE [42]. and revealed by western blotting using the antibodies described above. Western blotting, antibodies and immunofluorescence staining Silver staining and mass spectrometry Proteins were resolved on standard SDS-PAGE gels and After protein separation by SDS-PAGE, gels were fixed with 2 transferred to nitrocellulose membranes. Membranes were pre- washes in 40% methanol/10% acetic acid, sensitized in 30% blocked in 3% non-fat milk in PBS prior to addition of primary methanol, 0.2% sodium thiosulphate, 6.8% sodium acetate, antibodies. For Western blotting, coimmunoprecipitations, and impregnated with 0.25% silver nitrate and developed in 2.5% immunofluorescence, the following antibodies were used: anti- sodium carbonate, 0.15% formaldehyde, 0.003% sodium thiosul- PRAME (ab32185, Abcam WB 1:500, IF 1:100), anti-FLAG M2 phate. The staining reaction was stopped in 1.46% EDTA. (F3165, Sigma-Aldrich WB 1:500, IF 1:100), anti-Golgi 58K Bands of interest were excised and analysed by the Cambridge (ab27043, Abcam IF 1:100), anti-His (05–531, Millipore WB Centre for Proteomics. This involved overnight trypsin digestion of 1:1000), anti-ELB (sc11447, Santa Cruz Biotechnologies WB the proteins in the gel bands followed by reverse phase liquid- 1:200, IF 1:100), anti-ELC (sc1559, Santa Cruz Biotechnologies chromatography and then high mass accuracy mass spectrometry. WB 1:500, IF 1:10), anti-CUL2 (ab1870, Abcam WB 1:500, IF For identification of proteins, the MS/MS fragmentation data 1:250), anti-HA (3F10, Roche WB 1:500), anti-H2A (ab18255, were used to search the human National Centre Biological Abcam WB 1:500), anti-H2B (ab1790, Abcam WB 1:500), anti-H3 Information database using the MASCOT search engine. (ab1791, Abcam WB 1:1000), anti-H4 (ab7311, Abcam WB Probability-based Mowse scores were used for evaluating peptide 1:500), anti-GST (Ab-2, Oncogene Research Products WB 1:500), identifications [43]. Only those peptide matches with a Mowse anti-GFP (ab6556 Abcam WB 1:500). The following horse radish score .38 were considered significant and reported. peroxidise-linked secondary antibodies were used for Western blotting and co-immunoprecipitations at 1:5000: chicken anti- Yeast-two hybrid assays mouse (sc2954, Santa Cruz Biotechnologies), donkey anti-goat Yeast transformations, yeast two-hybrid reporter assays, yeast (30220–210, Alpha Diagnostics International), goat anti-rabbit cell-free extract preparations for western blotting were performed (sc2004, Santa Cruz Biotechnologies). For immunofluorescence, essentially as described previously [35]. the following secondary antibodies were used at 1:500: alexa fluor chicken anti-mouse 488 (A21200), alexa fluor chicken anti-goat Acknowledgments 594 (A21468) and alexa fluor chicken anti-rabbit 594 (A21442) Invitrogen. Y2H fusion proteins in yeast cell-free extracts were We thank Kunitoshi Yamanaka and Phillippe Gailly for expression vectors detected using anti-LexA (06–719, Millipore WB 1:1000) or anti- and reagents, and the Cambridge Protein Analysis facility for mass HA (3F10, Roche WB 1:500). spectrometry. We also thank Donald and Ada Olins for advice on Immunofluorescence staining and image capture was performed subcellular structures in HL60 cells. Special thanks to undergraduate project students Vee Mae Saw and Hazirah Azanam Nur for assistance in using a Leica LSM510 confocal microscope essentially as replicating qPCR data. described previously [39]. U2OS, HEK293 and MCF-7 cells were grown on glass coverslips, while non-adherent cells were mounted on glass slides at 1000rpm for 5 minutes using a

PLOS ONE | www.plosone.org 9 February 2013 | Volume 8 | Issue 2 | e58052 Function and Regulation of PRAME

Author Contributions DMH. Contributed reagents/materials/analysis tools: FHF JF AB NT KS. Wrote the paper: FRW DMH. Conceived and designed the experiments: FRW DMH. Performed the experiments: FRW JF HMC NT. Analyzed the data: FRW JF HMC KS

References 1. Wadelin F, Fulton J, McEwan PA, Spriggs KA, Emsley J, et al. (2010) Leucine- and high concentration of lipopolysaccharide is involved in NF-kappa b rich repeat protein PRAME: expression, potential functions and clinical activation. Infect Immun 69: 2788–2796. implications for leukaemia. Mol Cancer 9: 226. 22. Yuan Z, Teasdale RD (2002) Prediction of Golgi Type II membrane proteins 2. Ikeda H, Lethe B, Lehmann F, van Baren N, Baurain JF, et al. (1997) based on their transmembrane domains. Bioinformatics 18: 1109–1115. Characterization of an antigen that is recognized on a melanoma showing 23. Perkins DN, Pappin DJ, Creasy DM, Cottrell JS (1999) Probability-based partial HLA loss by CTL expressing an NK inhibitory receptor. Immunity 6: protein identification by searching sequence databases using mass spectrometry 199–208. data. Electrophoresis 20: 3551–3567. 3. Atanackovic D, Luetkens T, Kloth B, Fuchs G, Cao Y, et al. (2011) Cancer-testis 24. Shilatifard A, Conaway RC, Conaway JW (2003) The RNA polymerase II antigen expression and its epigenetic modulation in acute myeloid leukemia. elongation complex. Annu Rev Biochem 72: 693–715. Am J Hematol 86: 918–922. 25. Singer JD, Gurian-West M, Clurman B, Roberts JM (1999) Cullin-3 targets 4. Kessler JH, Beekman NJ, Bres-Vloemans SA, Verdijk P, van Veelen PA, et al. cyclin E for ubiquitination and controls S phase in mammalian cells. Genes Dev (2001) Efficient identification of novel HLA-A(*)0201-presented cytotoxic T 13: 2375–2387. lymphocyte epitopes in the widely expressed tumor antigen PRAME by 26. Litterman N, Ikeuchi Y, Gallardo G, O’Connell BC, Sowa ME, et al. (2011) An proteasome-mediated digestion analysis. J Exp Med 193: 73–88. OBSL1-Cul7Fbxw8 ubiquitin ligase signaling mechanism regulates Golgi 5. Matsushita M, Ikeda H, Kizaki M, Okamoto S, Ogasawara M, et al. (2001) morphology and dendrite patterning. PLoS Biol 9: e1001060. Quantitative monitoring of the PRAME gene for the detection of minimal 27. Schroder K, Sweet MJ, Hume DA (2006)Signalintegrationbetween residual disease in leukaemia. Br J Haematol 112: 916–926. IFNgamma and TLR signalling pathways in macrophages. Immunobiology 6. Paydas S, Tanriverdi K, Yavuz S, Disel U, Baslamisli F, et al. (2005) PRAME mRNA levels in cases with acute leukemia: clinical importance and future 211: 511–524. prospects. Am J Hematol 79: 257–261. 28. Starczynowski DT, Karsan A (2010) Innate immune signaling in the 7. Paydas S, Tanriverdi K, Yavuz S, Seydaoglu G (2007) PRAME mRNA levels in myelodysplastic syndromes. Hematol Oncol Clin North Am 24: 343–359. cases with chronic leukemia: Clinical importance and review of the literature. 29. Wolska A, Lech-Maranda E, Robak T (2009) Toll-like receptors and their role in Leuk Res 31: 365–369. hematologic malignancies. Curr Mol Med 9: 324–335. 8. Qin Y, Zhu H, Jiang B, Li J, Lu X, et al. (2009) Expression patterns of WT1 and 30. Lee SA, Kim SM, Son YH, Lee CW, Chung SW, et al. (2011) Peptidoglycan PRAME in acute myeloid leukemia patients and their usefulness for monitoring enhances secretion of monocyte chemoattractants via multiple signaling minimal residual disease. Leuk Res 33: 384–390. pathways. Biochem Biophys Res Commun 408: 132–138. 9. Quintarelli C, Dotti G, Hasan ST, De Angelis B, Hoyos V, et al. (2011) High- 31. Grivennikov SI, Greten FR, Karin M (2010) Immunity, inflammation, and avidity cytotoxic T lymphocytes specific for a new PRAME-derived peptide can cancer. Cell 140: 883–899. target leukemic and leukemic-precursor cells. Blood 117: 3353–3362. 32. Olins AL, Olins DE (2004) Cytoskeletal influences on nuclear shape in 10. Quintarelli C, Dotti G, De Angelis B, Hoyos V, Mims M, et al. (2008) Cytotoxic granulocytic HL-60 cells. BMC Cell Biol 5: 30. T lymphocytes directed to the preferentially expressed antigen of melanoma 33. Doolan P, Clynes M, Kennedy S, Mehta JP, Crown J, et al. (2008) Prevalence (PRAME) target chronic myeloid leukemia. Blood 112: 1876–1885. and prognostic and predictive relevance of PRAME in breast cancer. Breast 11. Steinbach D, Schramm A, Eggert A, Onda M, Dawczynski K, et al. (2006) Cancer Res Treat 109: 359–365. Identification of a set of seven genes for the monitoring of minimal residual 34. Epping MT, Hart AA, Glas AM, Krijgsman O, Bernards R (2008) PRAME disease in pediatric acute myeloid leukemia. Clin Cancer Res 12: 2434–2441. expression and clinical outcome of breast cancer. Br J Cancer 99: 398–403. 12. Tajeddine N, Millard I, Gailly P, Gala JL (2006) Real-time RT-PCR 35. Le Douarin B, Heery DM, Gaudon C, vom Baur E, Losson R (2001) Yeast two- quantification of PRAME gene expression for monitoring minimal residual hybrid screening for proteins that interact with nuclear hormone receptors. disease in acute myeloblastic leukaemia. Clin Chem Lab Med 44: 548–555. Methods Mol Biol 176: 227–248. 13. Ortmann CA, Eisele L, Nuckel H, Klein-Hitpass L, Fuhrer A, et al. (2008) 36. Harreman M, Taschner M, Sigurdsson S, Anindya R, Reid J, et al. (2009) Aberrant hypomethylation of the cancer-testis antigen PRAME correlates with Distinct ubiquitin ligases act sequentially for RNA polymerase II polyubiquityla- PRAME expression in acute myeloid leukemia. Ann Hematol 87: 809–818. tion. Proc Natl Acad Sci U S A 106: 20705–20710. 14. Roman-Gomez J, Jimenez-Velasco A, Agirre X, Castillejo JA, Navarro G, et al. 37. Li XS, Trojer P, Matsumura T, Treisman JE, Tanese N (2010) Mammalian (2007) Epigenetic regulation of PRAME gene in chronic myeloid leukemia. Leuk SWI/SNF – a subunit BAF250/ARID1 is an E3 ubiquitin ligase that targets Res 31: 1521–1528. histone H2B. Mol Cell Biol 30: 1673–1688. 15. Birtle Z, Goodstadt L, Ponting C (2005) Duplication and positive selection 38. Sasagawa Y, Otani M, Higashitani N, Higashitani A, Sato K, et al. (2009) among hominin-specific PRAME genes. BMC Genomics 6: 120. Caenorhabditis elegans p97 controls germline-specific sex determination by 16. Williams JM, Chen GC, Zhu L, Rest RF (1998) Using the yeast two-hybrid controlling the TRA-1 level in a CUL-2-dependent manner. J Cell Sci 122: system to identify human epithelial cell proteins that bind gonococcal Opa 3663–3672. proteins: intracellular gonococci bind pyruvate kinase via their Opa proteins and require host pyruvate for growth. Mol Microbiol 27: 171–186. 39. Kindle KB, Troke PJ, Collins HM, Matsuda S, Bossi D, et al. (2005) MOZ-TIF2 17. Epping MT, Wang L, Edel MJ, Carlee L, Hernandez M, et al. (2005) The inhibits transcription by nuclear receptors and p53 by impairment of CBP human tumor antigen PRAME is a dominant repressor of retinoic acid receptor function. Mol Cell Biol 25: 988–1002. signaling. Cell 122: 835–847. 40. Johannes G, Carter MS, Eisen MB, Brown PO, Sarnow P (1999) Identification 18. Costessi A, Mahrour N, Tijchon E, Stunnenberg R, Stoel MA, et al. (2011) The of eukaryotic mRNAs that are translated at reduced cap binding complex eIF4F tumour antigen PRAME is a subunit of a Cul2 ubiquitin ligase and associates concentrations using a cDNA microarray. Proc Natl Acad Sci U S A 96: 13118– with active NFY promoters. EMBO J 30: 3786–3798. 13123. 19. Tajeddine N, Gala JL, Louis M, Van Schoor M, Tombal B, et al. (2005) Tumor- 41. Johannes G, Sarnow P (1998) Cap-independent polysomal association of natural associated antigen preferentially expressed antigen of melanoma (PRAME) mRNAs encoding c-myc, BiP, and eIF4G conferred by internal ribosome entry induces caspase-independent cell death in vitro and reduces tumorigenicity in sites. RNA 4: 1500–1513. vivo. Cancer Res 65: 7348–7355. 42. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using 20. Tian X, Pascal G, Monget P (2009) Evolution and functional divergence of real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: NLRP genes in mammalian reproductive systems. BMC Evol Biol 9: 202. 402–408. 21. Liu Y, Wang Y, Yamakuchi M, Isowaki S, Nagata E, et al. (2001) Upregulation 43. Pappin DJ, Hojrup P, Bleasby AJ (1993) Rapid identification of proteins by of toll-like receptor 2 gene expression in macrophage response to peptidoglycan peptide-mass fingerprinting. Curr Biol 3: 327–332.

PLOS ONE | www.plosone.org 10 February 2013 | Volume 8 | Issue 2 | e58052